Skip to content

TaiwanVVM-style simulations

High-resolution simulations over Taiwan use realistic topography and land-surface fields read from NetCDF, together with the Noah land model when enabled. This workflow is often called TaiwanVVM in project documentation.

Configuration overview

  1. Grid and domain — Set grid.nx, grid.ny, grid.nz, and spacing to match your experiment. fix_lonlat: true is typical for lon/lat-consistent setups (see sample default_config.json).

  2. Spatial input file — netcdf_reader.source_file must point to a NetCDF file containing the variables you need. The sample default_config.json lists 2D variables such as lon, lat, topo, sea_land_ice_mask, vegetation and soil types, slopes, ground temperature, albedo, greenness, LAI, etc. Your file must provide the same names (or you must change the list and preprocessing to match).

  3. Land and surface — Enable physics.land.enable_land for Noah; tune frequency_step. Enable physics.surface.enable_surface and choose mode (e.g. sflux_tc_2d) when you need surface flux coupling. These interact with the 2D fields read from NetCDF.

  4. Output — Set output.output_dir to a writable path; list diagnostics you need in output.fields_to_output (including topo, lon, lat if you want them in the archive).

Generating spatial NetCDF

The script scripts/generate_init_nc.py reads rundata/input_configs/default_config.json (same path as the main config used in the script) and writes:

  • netcdf_reader.source_file from the config (directories are created if missing).

It supports:

  • USE_TAIWAN_TOPO = True — Reads high-resolution Taiwan data from SOURCE_TW_DATA (default ../rundata/land/topolsm_TW.nc) and coarsens to the configured grid.
  • USE_TAIWAN_TOPO = False — Idealized ridge and land-type patterns for synthetic experiments.

Edit the constants at the top of the script and run it from the scripts/ directory with Python dependencies (netCDF4, numpy, scipy) installed. Align paths with your machine.

Large-scale forcing (optional)

For lateral boundary forcing, scripts/generate_ls_forcing.py can prepare forcing files; dynamics.forcings.lateral_boundary_nudging in JSON points to directories and file naming under rundata/LS_forcings/ when enabled.

Operational notes

  • Ensure NetCDF and PnetCDF libraries used at build time match the files you read and write.
  • Taiwan runs are often large; adjust MPI ranks, physics.rrtmgp.column_chunk_size, and I/O strategy (output.engine, --io-tasks) for your cluster.
  • The repository README credits Noah land GPU work to the Central Weather Administration (CWA) of Taiwan.

For a minimal first run, start from rundata/input_configs/default_config.json, replace output.output_dir and initial_conditions.source_file with local paths, and verify netcdf_reader.source_file exists after preprocessing.